Thallium-205 nuclear magnetic resonance study of thallium(III) halide

Wanzhi Chen, Fenghui Liu, Kazuko Matsumoto, Jochen Autschbach, Boris Le Guennic, Tom Ziegler, Mikhail Maliarik, and Julius Glaser. Inorganic Chemistry...
0 downloads 0 Views 982KB Size
6642

J. Am. Chem. SOC.1981, 103, 6642-6649

18-electron intermediates (e.g., HRh(P-P)(P-P)m(olefin) or HRh(P-P)(PR,)(olefin)) that have three phosphine ligands coordinated to Rh at the step in the mechanism where Rh-H addition to coordinated olefin occurs (Le., the selectivity determining step).

Acknowledgment. We gratefully acknowledge helpful discussions with Dr. J. Unruh and the assistance of R. Carney and H. Pate1 in preparation of materials and of Mrs. J. Vieira for obtaining NMR spectra. We also thank one of the reviewers for suggesting the IR work and Dr. J. Rafalko for conducting it.

Thallium-205 Nuclear Magnetic Resonance Study of Thallium( 111) Halide Complexes in Aqueous Solutions Julius GIaser*ln and Ulf Henrikssonlb Contribution from the Departments of Inorganic Chemistry and Physical Chemistry, The Royal Institute of Technology, S-100 44 Stockholm 70, Sweden. Received January 5, 1981

Abstract: A combination of solution and solid-state *OST1NMR experiments is used to study the formation and geometry of TIX,'" complexes (X = C1, Br). The chemical shifts for the individual species and their stability constants are determined for dilute (0.05 M) and concentrated (1.0 and 2.6 M) aqueous TI(II1) solutions. The existence of TIC]:- and TIC1:- as well as at least one species higher than TIBrc in solution is shown. The species TIC], is 300-400 ppm less shielded in aqueous solutionsthan in solid phase, indicating a structural difference. The chemical shifts obtained for the individual TIXnf" complexes are compared with corresponding shifts for zinc(I1) and cadmium(I1) halide complexes.

The complexes formed in the thallium(II1) halide system are among the strongest metal ion halide complexes known. Several investigations were made on this and some interesting structural properties can be discerned. Thus, for example, there are indication^'^^^ that in aqueous solutions of T1(C104), only two water molecules are strongly bound to T13+and that these are replaced by chloride when T1C1(H20),2' and TIC12(H20)4+are formed. The stability constants and the enthalpy va1ues1s13J6,20 indicate some structural change between the second and third complex. The existence of higher complexes TICI,'-n ( n > 4) was for a long time a matter of controversy. In dilute solutions, where all of the emf, solubility, and calorimetric measurements were per(1). (a) Department of Inorganic Chemistry. (b) Department of Physical Chemistry. (2) Lee, A. G. "The Chemistry of Thallium"; Elsevier: Amsterdam, 1971; pp 44-91 and references cited therein. (3) Benoit, R. Bull. SOC.Chim. Fr. 1949, 518. (4) Peschanski, D.; Valladas-Dubois, S. Bull. SOC.Chim. Fr. 1956, 1170. (5) Horrocks, D. L.; Voigt, A. F. J . Am. Chem. SOC.1957, 79, 2440. (6) Figgis, B. N. Trans. Faraday SOC.1959, 55, 1075. (7) Freeman, R.; Gasser, R. P. H.; Richards, R. E. Mol. Phys. 1959, 2, 301. (8) Baysal, B. Actes Congr. Int. Catal., 2nd 1960, I , 559 (Chem. Abstr. 1962, 56, 1140~). (9) Busev, A. I.; Tiptsova, V. G.; Sokolova, T. A. Vestnik Moskou. Uniu. Khim. 1960.11-6, 42. (10) Ahrland, S.; Grenthe, I.; Johansson, L.; Noren, B. Acta Chem. Scand. 1963, 17, 1567. (11) Ahrland, S . ; Johansson, L.Acta Chem. Scand. 1964, 18, 2125. (12) Woods, M. J. M.; Gallagher, P. K.; Hugus, 2. 2.;King, E. L.Inorg. Chem. 1964, 3, 1313. (13) Leden, I.; Ryhl, T. Acta Chem. Scand. 1964, 18, 1196. (14) Nord., G.; Ulstrup, J. Acta Chem. Scand. 1964, 18, 307. (15) Spiro, T. G. Inorg. Chem. 1965, 4 , 731. (16) Kul'ba, F. Y.; Mironov, V. E.; Mavrin, I. F. Zh. Fiz. Khim. 1965,39, 2595. (17) Yakovlev, Y. B.; Kul'ba, F. Y.; Mironov, V. E. Zh. Neorg. Khim. 1967, 12, 3283. (18) Davies, E. D.; Long, D. A. J . Chem. SOC.A 1968, 2050. (19) Schmidt, K. J . Inorg. Nucl. Chem. 1970, 32, 3549. (20) Biedermann, G.; Spiro, T. G. Chem. Scr. 1971, I , 155.

0002-7863/8l/l503-6642$01.25/0

formed, the formation of higher complexes is generally claimed to be negligible: K5 = [T1X5*-]/( [TlXd-] [X-1) < 0.07 M-' for X = C1 and K5 < 0.37 M-' for X = Br, at 25 OC.Io,ll (The exceptions are some emf works, where silver halide electrodes have been used.3-4*8g9 These results are probably erroneous because of the oxidation of the electrodes by Tl(III), as was pointed out by some investigators.lO) However, UV spectra of dilute Tl(II1) solutions at very high chloride concentration^'^ indicate the formation of one higher species, TlCl:-, with the stability constant K5 = 0.8 (2) M-I. Extraction data for dilute solutions are also interpreted in terms of weak bonding of the fifth chloride ligand (Ks= 0.3 M-'),5 although other workers using the same method found no evidence for complexes higher than TlC14- (at least not more than 3% of the total thallium c~ncentration).'~ In concentrated solutions, Raman spectra clearly show that, in the chloride system, at least one higher complex exists, probably TQ3- (K4.6 = [T1C1:-]/([TlC14-] [Cl-1') = 0.2 M-2).'5*'8 N M R data also indicate the formation of complexes with more than four chlorides per Tl(III).6*7 Figgis6 postulated that the most stable species a t higher halide concentrations is TI2Xg3-(X = C1, Br). For determination of the compositions and structures of the complexes formed in aqueous solutions containing Tl(II1) and one of the halides (C1, Br), an X-ray diffraction study2' was undertaken. For the interpretation of the diffraction data the distribution of TI(II1) among the different species has to be known for the concentrated solutions ([TI], 1 1 M). However, all the available stability constants are determined for rather dilute solutions I50 mM). The aim of the present work is to estimate this distribution. In addition, the 205Tlchemical shifts for the individual TlX:-" complexes are obtained, and they may provide some structural information when compared with the chemical shifts for the corresponding solid-state coordination compounds with known structures. The shifts can also be compared with those determined for the halide complexes of the dl0 ions ZnZ+and Cd2+. (21) Glaser, J.; Johansson, G. Acta Chem. Scand., in press.

0 1981 American Chemical Society

J . Am. Chem. SOC.,Vol. 103, No. 22, 1981 6643

Study of Thallium(III)Halide Complexes Table I.

Chemical Shifts (6) and Line Widths at Half-Height (Au, ,,) for Various Solid TKIII) Salts‘ compound

6

2175 2051 27 16 2022 1972 2007 2019 1926 1098 1262 -1271 -1194 -1560

Tl(C10,),~6Hz0 TlCl, . 4 H z 0 KTlCI, C s,T1ClS .H ,0 Na,TICl, .12H,O K,TlC1,.(1 3/7)H,O [CoWH,), ITlC1, Cs,TI,Cl, TlBr .4H ,0 KT1Br,.2H20 [Co(NH,), ITlBr, Cs,TI,Br, [NBu,]TlI,

,

Aui/23

kHz

configuration around Tl(II1) Tl(H,0),3+ regular o c t a h e d r ~ n ~ ~ TICl,(H 20)z trigonal b i ~ y r a m i d e ~ ~ ? ~ ’ T U 4 - tetrahedr~n~’,~’ T1Cl,(H,0)2‘ ~ c t a h e d r o n ~ ~ . ~ ’ TlCI,,- octahedron,, TlCI,,- octahedron and T1C1s(H,0)2‘ ~ctahedron’’*~’, TlCI,,- o ~ t a h e d r o n ~ ~ ~ ~ ’ TlCI,’- unit (two TlC16,- octahedra with one common TlBr,(H,O), trigonal b i p y r a ~ n i d e ~ ~ , ~ ’ TIBr,- t e t ~ a h e d r o n ~ ~ TlBr,,‘ o ~ t a h e d r o n , ~ unknown slightly distorted T11,- tetrahedrons3

6.8 8.0 6.0 5.1 5 .O 6.0 3.3 9.0 14.0 15.5 19.0 12.0 20

The shifts are given in ppm toward higher frequency with respect to Tl+(aq) at infinite dilution,

The following measurements were performed. The 20ST1 chemical shifts for 50 mM, 1.0 M, and 2.6 M aqueous solutions of Tl(II1) and varying concentrations of the halide were measured. In the 50 mM solutions, separate signals were obtained for the lower complexes (n = 0, 1 for C1; n = 0, 1, 2 for Br), due to slow exchange. In all the other solutions only one average signal was observed due to rapid exchange. The position of this signal is given by 6 p d

= Cp;b, n

(1)

where p,’ = [TIXn~n]//[Tl], and 6, is the chemical shift of TIXnsn.

Experimental Section Materials. Natural abundance TI chemicals were used throughout. The preparation of the solid phases was described elsewhere.2632 The investigated solutions were prepared from stock solutions: 2.1 M TI(CIO& in 2.65 M HC104,31~33 2.6 M T1X,(C104)3-n,3’ concentrated HC104 (p.a., free from halide), 8.4 M NaC104,33and varying amounts of solid LiX (p.a.). They were analyzed for TI (bromatometric titratiodk with potentiometric end point), halide (Volhard titration3”), and [H’] (titration with NaOH after reduction of TI3+ to TI’ by H20;3). All bromide solutions witkR,, = [Br-],,/[TI], I 2 were saturated with Br2 in order to prevent the reduction of TI(II1) to Tl(1). The concentration of TI(1) was lower than 1% of the total thallium in all the solutions investigated. Measurements. The 5 1.9 MHz 20sTlspectra were recorded at the ambient probe temperature of 27.2 OC by using a Bruker CXP-100 spectrometer. The field was stabilized with an external field lock. The samples were contained in 10-mm nonspinning sample tubes. The contribution to the line width from field inhomogeneities was 18 Hz. All 205T1chemical shifts were reported in parts per million toward higher frequency with respect to an aqueous solution of TICIO, at infinite dilution. Since the chemical shifts for aqueous solutions of different TI(1) salts extrapolated to the same value at infinite diluti~n,’~ this value corresponded to the chemical shift of the free hydrated TI+ ion. The accuracy of the chemical shifts determined depended strongly on the width of the observed peaks. For narrow lines it was better than 0.2 (22) Maciel, G. E.; Simeral, L.; Ackerman, J. J. H. J . Phys. Chem. 1977, 81, 263. (23) Drakenberg, T.; Bjsrk, N. 0.;Portanova, R. J. Phys. Chem. 1978, 82, 2423. (24) Ackerman, J. H. H.; Orr, T. V.; Bartuska, V. J.; Maciel, G. E. J . Am. Chem. SOC.1979,101, 341. (25) Ingn, N.; Kakolowicz, W.; SillBn, L. G.; Warnqvist, B. Talanta 1967, 14, 1261. (26) Pratt, J. H. Am. J . Sei. 1895, 49, 397. (27) Meyer, R. J. Z . Anorg. Allg. Chem. 1900, 24, 321. (28) Hoard, J. L.; Goldstein, L. J . Chem. Phys. 1935, 3, 645. (29) Watanabe, T.; Atoji, M.; Okazaki, Ch. Acta Crystallogr. 1950, 3, 405. (30) Andrews, S. P.; Badger, P. E. R.; Goggin, P. L.; Hurst, N. W.; Rattray, J. M. J . Chem. Res. Miniprint 1978, 1401. (31) Glaser, J. Thesis. Royal Institute of Technology, S-100 44 Stockholm, 1981. (32) Glaser, J.; Johansson, G. Acta Chem. Stand., in press. (33) Biedermann, G. Ark. Kemi 1953, 5, 441. (34) (a) Noyes, A. A.; Hoard, J. L.; Pitzer, K. S . J. Am. Chem. SOC.1935, 57, 1231. (b) Kolthoff, I. M.; Sandell, E. B. “Textbook of Quantitative Inorganic Analysis”; Macmillan: New York, 1961; p 546. (35) Dechter, J. J.; Zink, J. I. J . Am. Chem. SOC.1975, 97, 2937.

Table 11. ”’Tl Chemical Shifts (6) and Line Widths at Half-Height ( A u , , , ) for Various Solution? ~~

solution

0.05 M T13+in conct HC1 0.05 M T13+in conct HBr 0.05 M T13+in conct HClO, 0.5 M [N(C,H,),]TlI, mother liquor from Tl(CIO,), .6H,O(s)

solvent H,O H,O H,O CH,Cl, HZO

6

A U , ~ kHz ,,

2096 1169 2056 -1732 1944 (90%) 1965 (10%)

0.03 0.01 0.03 0.2 0.6 0.2

a The shifts are given in ppm toward higher frequency with respect to Tl+(aq) at infinite dilution.

ppm. For the broadest lines the errors were considerably greater. No corrections were made for differences in bulk susceptibilities (they are probably less than 1 0 . 5 ppmZZand their influence on the results can therefore be neglected).

Results Solids. The 205Tlchemical shifts and line widths for the salts investigated are given in Table I. As can be seen, the lines are very broad, especially for the bromides. Hence, the accuracy in the determined shifts is rather low (-*25 ppm for the chlorides and -*50 ppm for the bromides). It was not possible to obtain any information about t h e individual shielding tensor elements from the solid-state spectra. In some cases our results differ significantly from earlier reported One possible reason for this difference is that it is often difficult to prepare solid samples that are completely free from mother liquor, and the narrow signal from the mother liquor is relatively easy to observe (see, e.g., Figure 2). For most of the solid samples we have recorded the signal both from the salt and from the mother liquor to assure that no erroneous assignments are made. Unfortunately, there is no literature data available on any salt containing the fifth bromide complex T1Br5(H20)2- alone. The salt (N,N’-dimethyltriethylenediammonium)pentabromothallate, (DMTEA)TlBr5, contains highly asymmetric T1BrSZ-units probably because of the packing in the ~ r y s t a l , ~and ’ this kind of complex is not expected to occur in aqueous solution. The only solid, in which the hydrated complex T1Br5(H,O)” has been found, But, as only 2 of 14 thallium atoms is Rb3TlBr6.(’3/7)H20.3’.38 a r e coordinated in this way, it was not possible to obtain the shift for the fifth complex. In fact, no resonance was found for this compound, probably because of the extreme broadening of the line.

Solutions. Three series of solutions were investigated. Within each series, the total thallium concentration was kept constant (0.05, 1.0, and 2.6 M), and the halide concentration was varied. In the 0.05 M solutions, the ionic medium was exactly the same as in ref 10 and 11 ( 3 M HC104 + 1 M NaC10,) in order to maintain the validity of the equilibrium constants. In the 1.0 M (36) (a) Hafner, S.; Nachtrieb, N. H. J . Chem. Phys. 1964,40,2891. (b) Schneider, W. G.; Buckingham, A. D. Discuss. Faraday Soc. 1962, 34, 147. (37) Shriver, D. F.; Wharf, I. Inorg. Chem. 1969, 8, 2167. (38) Glaser, J., to be submitted for publication.

6644 J. Am. Chem. SOC.,Vol. 103, No. 22, 1981

Glaser and Henriksson

TIC1",-"

r _

0.05 M 2000

k

1 6

15001

r

4

1000-

0

1

2

3

4

5

6

7

8

Q

1 0 1 1 1 2

I 4 I bI#*,

0

1

2

3

4

5

6

7

b -1 / [TI1

8

9

1011

12

101

Figure 1. 205TIchemical shifts from aqueous solutions of TI(II1) as functions of Rx = [X-],/[T1],, (X = CI, Br). 0 indicates experimental points, full lines are calculated from eq 1 by using the stability constants and chemical shifts for individual complexes given in Table 111. The fraction of TI(II1) present as TlX:-", calculated by using the same stability constants, is also shown. Separate signals for the different complexes were observed for the Rx values for which two experimental points are marked in the figure.

solutions, the sum [H+],,, + [Li'],,, = 3.0 M was kept constant for R = [X-]lo,/[Tl]l,,, I6.0. For [TI(III)] = 2.6 M,no ionic medium was used (moreover, [Tl],,, was 15% lower for R I 1). These conditions were chosen since the concentrated solutions should have exactly the same composition as those investigated by X-ray d i f f r a c t i ~ n . ~ ~ . ~ ~

-

The measured chemical shifts are shown in Figure 1. Some additional shifts are presented in Table 11. For the 0.05 M solutions separate signals were observed for the species TI3+, T1CI2+,TIBr2+,and TlBr2+, showing that the rate of exchange between these complexes is slow on a time scale defined by the difference between their chemical shifts. The signals are, however,

Study of Thallium(III) Halide Complexes

J . Am. Chem. SOC.,Vol. 103, No. 22, 1981 6645

Table 111. Calculated loST1Chemical Shiftsa and Stability Constants (Recalculated to 25 "C) for the TlX,'-,

[Tl]t,t = 0.05 M, [NaClO,] = 1 M, [HClO,] = 3 M

[ n l t o t = 1.0 M, + [Li+]= 3 M

[Tl]tot = 2.6 M

[HCIO,]

solids

Complexesb [Tl]tot G 0.05 M C [NaCIO,] = 1 M [HClO,] = 3 M

X=Cl 6 0

61

6, 6 3

6,

2086 (51) 2198 (*4) 2201 (2) 2412 (5) 2645 (2)

2075 (4) 2209 (4) 2183 (5) 2457 (9) 2630 (3)

2048 (15) 2194 (9) 2220 (10) 2357 (12) 2649 (6)

1.9 (5) x 3.6 (5) X 3.1 (1) x 1.0 (3) x

1.5 (4) x 107 0.6 (2) x 1013 5.0 (7) x 10'6 4.8 (2) x 1019 1.4 (2) x 1019 1.1 (2) x 1019

3 (1) x 10' 1.8 (6) X 10" 6.8 (9) X l o t 6 3.8 (3) x 1019

6s 66

01 P Z

03 04

P,

1013 1OI6 1019

1019

06

2.9 (3)

2175 (530) 2051 (t35) 2716 (530) 2022 ( t 2 0 ) 1972 (r25) 2.9 (t0.3) x 10' 1.8 (i0.2) x 1013 4.5 (50.5) x 1 O I 6 2.8 (*0.4) x 4) was negligible; the parameters for the first five complexes ( n = 0-4) could be refined. (2) All the data were used in the refinement, but some or all parameters for the lower complexes were kept constant at the values from step 1. In some cases, when all the parameters could not be refined simultaneously, they were refined in subgroups, in different combinations, until no further change was obtained. The standard deviations given in Table I11 were increased to account for the variation of the parameters during the different refinements. However, it should be emphasized that in the least-squares refinements the experimental points close to the domination regions of the different TlX,,-" complexes are weighted up as a consequence of the calculation procedure used. Thus the standard deviations in the stability constants for the concentrated solutions are valid only in these regions and are probably much higher outside the domination areas as a result of the dependence of the stability constants on the ionic strength. The reliability of the simultaneous determination of stability constants and chemical shifts for the different species can be tested for the 50 mM solutions. It is seen from the results in Table 111 that our stability constants agree with those of Ahrland et al.lOgll within the standard deviations. The situation is in fact quite favorable for the thallium(II1) halide complexes since (1) each TlX,,-" complex has a predominance region for some halide to thallium ratio (cf. distribution curves in Figure 1, calculated by the HALTAFALL program2s), (2) the chemical shift range for the TlX,," complexes in aqueous solutions is very wide, almost 2000 ppm, and the shifts can be determined quite accurately, and (3) the high NMR sensitivity of 20sTlmakes it possible to record the spectra at a low total thallium concentration, 50 mM, where the results can be compared with accurate stability The method has, however, some limitations. When several complexes exist in solution, the change of the ionic strength (in the concentrated solutions) may alter the distribution of the complexes (40) Heistand, R. N.; Clearfield, A. J . Am. Chem. SOC.1963, 85, 2566. (41) (a) Br~nsted,J. N . J . Am. Chem. SOC.1922,44, 877. (b) Guggenheim, E. A. "Application of Statistical Mechanics";Clarendon Press: Oxford, 1966. (c) Scatchard, G. Chem. Rev. 1936,19,309; 1.Am. Chem. SOC.1968, 90, 3124. (d) Biedermann, G. In "The Nature of Seawater"; Goldberg, E. D., Ed.; Dahlem Konferenzen: Berlin, 1975; pp 339-362. (42) Biedermann, G.; Glaser, J., to be submitted for publication.

Table IV. Dependence of the Chemical Shiftsa for the Individual Complexes (6,,) on the Ionic Strength (I)and the Hydrogen Ion Concentration of the Solution

solution composition

[H+], I, mol. mol. L - ~ L-' complex

0.05 M T l C l o ~ s ( C 1 0 4 ) 2 ~ s 3

4

0.05 M TlCl,,s(C104)2,s

3

7

0.05 M TlBro.s(C104)2~s 3

4

0.05 M TlBro.s(C104)2.s

3

7

n3+

0.05 M TlBrl~s(C104)l.s 3

4

T1Br2+ TlBr" TlRr.'

0.05 M TlBr,.5(C104),.s

3

7

0.05 M Lil.sTlC14,s 1.0 M Li,.oTlC14.0 2.6 M Lil.aTlC14.a 0.05 M Lil,6TlBr4.6 1.0 M Lil,oTlBr4,a 2.7 M Lil,oTlBr4.0

3 2 0 3 2 0

4 3 2.6 4 3 2.7

T13+ TICl" Ti3+

TU2+ TI3+ T1Br2+

TlC1;TlC1,TlC1,T1Br4T1Br4TlBr,-

6,

2077 2195 2069 2190 2084 1539 2067 1513 1535 767 1502 747 2634 2625 2624 1317 1317 1319

The shifts are given in ppm toward higher frequency with respect to Tl+(aq) at infinite dilution.

and/or the 6,:s. The two effects may be difficult to separate. For the solutions where the chemical shifts for the individual complexes could be measured directly due to the slow exchange, the dependence of 6, on ionic strength was studied. Some of the results are presented in Table IV. The peaks from the complexes with n = 0, 1 and 2 are, however, rather broad (1-2 kHz). Hence, the uncertainty in the determined shifts can be estimated to approximately *7 ppm. As can be seen, the shifts for T13+,TlBr2+,and TlBr2+ show a small, but probably significant, change to lower frequency when the ionic strength is increased from 4 to 7 M. For the TlBr4- species, which accounts for more than 99% of the total thallium in solutions where R B ~ k 4, there is no significant change in the shift when the ionic media is varied. The same conclusion is probably valid for TlCl,, which accounts for -95% of the total thallium present for R a k 4. For n = 5 and 6 the complexes are comparatively weak, and it was not possible to obtain the 6,:s from the refinements since there is a strong correlation between the shifts and the stability constants. Therefore, for an estimation of the stability constants f15 and 66, the shifts from corresponding solid compounds were used. This procedure is reasonable because of the good agreement between the shifts from aqueous solutions and solids found for CdX; (X = C1, Br, I) and for most of the Tl(II1) complexes (vide infra). For the chlorides, both the fifth and the sixth complexes in the solid state are well-known and the structures established. For the bromides, only the sixth complex is known in the solid state. Thus, in order to use the solution data for R B =~ [Br-],,/[TlI,,, > 4, it was necessary to assume that only the sixth complex is formed and to use the shift for T1Br:from the solid state. This assumption may be plausible (at least for the concentrated solutions) when one takes into account that the fifth complex does not occur a t all in the solid state, whereas the sixth complex can be found in several salts which can be precipitated from aqueous solutions. The great width of the signal from the solid salt is not a serious problem for the estimation of the stability constant for T1Brs3~ ~ ppm. ~ ~ since the difference between 6nBr,-and 6 m is -2500 It is important to point out that the accuracy of our measurements for Rx > 4 is better than -0.2 ppm for all the investigated solutions. The differences between the experimental values and the shifts calculated by using an improper model are 2 orders of magnitude higher than the experimental accuracy. For all the data sets (Cl, Br), the reported stability constants for the fifth and sixth complexes do not change by more than two standard deviations when gs and 66 are changed by the highest possible error in their determined values &e., 30 ppm for the

J. Am. Chem. SOC.,Vol. 103, No. 22, 1981 6641

Study of Thallium(III)Halide Complexes

10

I\

0

1 0

,

~,

,

,

,

,

5

,

~,

,

,

, I

10

CI

Figure 3. The variation of the ionic strength with Rcl = [Cl-],ot/[T1],,, for the different total thallium concentrations.

chloride compounds and 70 ppm for the hexabromide). The 0.05 M Solutions. Separate lines were observed for the different complexes for [C1-]~ot/[T1],ot 1 and for [Br-]~ol/[T1],ot ~ ~ be ~ measured + directly. 2. Thus, bTIC12+, 6nBr2+, and 8 ~ j could The remaining shifts and stability constants were obtained by least-squares refinements of the data for higher halide/[Tl],,, ratios. The 1 M Solutions. The lines for the bromide system for R B r I 2 are very broad and were not used in the refinements. Therefore, it was not possible to determine PI and hl. The number of experimental points per determined parameter decreased, resulting in high standard deviations for b2, b3, b4, and probably a less reliable value of P2. The chloride system behaved well, and accordingly relatively low standard deviations were obtained (except for (3, and b,, which are discussed below). The 2.6 M Solutions. The limited solubility in both systems did not permit the investigation to be extended beyond RcI = 6.4 and R B=~ 5.0. The ionic strength in the solutions varies strongly, as shown in Figure 3. Consequently, as the shifts are somewhat sensitive to this variation, the standard deviations in the 6,:s are higher. For the chloride solutions, PI is not very well determined, since the dependence of BObsd on Rcl is weak in this region. Discussion Chemical Shifts. The agreement between the 6,:s for the different thallium concentrations is surprisingly good (with two exceptions discussed below), indicating that the assumption of their independence of ionic strength is reasonable. This fact can be understood if no change occurs in the structures of the complexes, when the ionic medium is changed. The comparison with corresponding solids with known structures may provide some additional information. The chemical shift difference (- 100 ppm) between 6nJ,9)3+ in the solid and in solution (cf. Table 111) is certainly significant. In T1(C104)3-6H20,a regular octahedral T1(H20),3+ unit has been found.32 In this salt, the water molecules are probably oriented trigonally (i.e., the negative sides of the dipoles point toward T13+) because of the packing in the crystal. This would also be in agreement with the fact that small three-charged metal ions are often coordinated in this way.43 Each T1(H20)63+octahedron is surrounded by 14 perchlorate groups. In aqueous solutions of this salt, on the other hand, the situation is partially changed. As shown in the recent X-ray diffraction investigation of thallic perchlorate solutions in water,21such solutions contain regular octhedral Tl(H20)63+units with the T1-0 distance not significantly different from the distance in the solid. However, in the solutions the concentration of free water is much higher than [C104-]. The water molecules also have a greater ability than the perchlorate ions to participate in hydrogen bonds. Therefore, the second

coordination sphere around the Tl(H20)63+octahedron probably consists mostly of water molecules. Furthermore, these solutions were prepared by using an excess of acid, in order to avoid precipitation of T1203. The high concentration of [H’] and the good accessibility of water molecules make it possible for some of the water molecules in Tl(H20)63+to be tetrahedrally oriented (Le., pointing toward the metal ion with one-electron pair, accepting a hydrogen bond from another water molecule, and donating two hydrogens bonds). As seen in Table 11, two rather narrow peaks are observed from the mother liquor in contact with Tl(C104)36H20(s). The signals disappear when perchloric acid is added. These signals may originate from hydrolysis complexes present in the mother liquor. The existence of such species in dilute aqueous solutions was established by Biedermanx~.~,Further discussion of these complexes is outside the scope of this work. The shift for T1Br2+in the 2.6 M solutions is 150 ppm lower than that in the dilute solutions. Even though the peaks from the concentrated solutions are very broad in this region and the ionic strength varies a lot, this difference may still be an indication of some change in geometry of the complex. The crystal structures of TlX3.4H20 (X = C1, Br) contain trigonal-bipyramidal TIX3(H20)2units with the water molecules in the apex positions and the thallium a little off the plane of the three halide^.^^,^^ The chemical shifts for the TlBr, complex in solution agrees rather well with the shift for the solid compound, indicating that the geometry of the TlBr3 species is approximately the same in both phases. Still, there is a difference of -90 ppm between the shifts in the solid and in the solutions which may be due to a small change in the position of T1 in the complex (more or less off the plane of the bromides) or may only reflect some difference in the second coordination sphere (e.g., different hydration). For the TlC1, complex the situation is quite different. First, there is a surprisingly large difference ( 30&400 ppm) between the shifts for the trigonal bipyramid in the solid state and for the complex in the solution. Second, 6nc13changes significantly with the total thallium concentration (cf. Table 111). This indicates that T1Cl3 in aqueous solution is not a trigonal bipyramid, but probably a tetrahedron, T1C13(H20). Such a tetrahedral geometry has been found for HgX3(H20)- (X = Br, I) in aqueous sohti on^.^^^ Another possibility is a coexistence of the trigonal bipyramid and the tetrahedron, as found for HgX, (X = C1, Br) in various organic solvents.46b The shifts of the fourth complexes in solids are very close to those calculated for the solutions, as these species are tetrahedral in both phases.21~31~47-5” X-ray powder photographs show that the compounds Cs3T12C19 and Cs3T12Br9are not isomorphous. Cs3T12C19is known to contain T12C1,3- units (two TlC16 octahedra sharing one and its chemical shift is rather close to that of other salts containing the TlC1:- complex. The shift of Cs3T12Br9is very close to that of the octahedral T1Brs3. Thus, it is probable that Cs3T12Br9contains a Tl2Br?- unit of the same type as found in Cs3T12C19. It is noteworthy that the 205Tlshifts for Tl14- (both in the solid and in CH2C12solution), Cs3T12Br9(s),and C O ( N H , ) ~ T ~ B ~ ~ ( S ) fall in the region -1200 to -1800 ppm where no T1 shifts were reported before.52 The shift from the solid [N(C4H9)4]T114differs almost 200 ppm from the shift of this salt in dichloromethane solution. This difference is experimentally significant and may be due to the different degree of distortion of the Tl14- unit in the two phases N

(46) (a) Sandstrom, M.; Johansson, G . Acta Chem. Scand 1977.31, 132. (b) Waters, D. N.; Kantarci, Z.; Rahman, N. N. J . Raman Spectrosc. 1978,

7, 288. (47) Thiele, G.; Grunvald, B.; Rink, W.; Breitinger, D. 2.Nuturforsch., B Anorg. Chem., Org. Chem. 1979, 348, 1512. (48) Glaser, J. Acta Chem. Scand. 1980, 34, 75. (49) Glaser, J. Acta Chem. Scand. 1980, 34, 157. ( 5 0., ) Hazell. A. C- . J... Chem. .. SOC.1963. -. --,-3459. (51) Powell, H. M.; Wells, A. F. J . Chem. SOC.1935, 1008. (52) Hinton, J. F.; Briggs, R. W. In “NMR and the Periodic Table”; Harris, R. K.; Mann, B. E., Eds.; Academic Press: London, 1978; pp 288-308. ~~

(43) Friedman, H. L.; Lewis, L. J . Solution Chem. 1976, 5, 445. (44) Zvonkova, Z. V. Zh. Fir. Khim. 1956, 30, 340. (45) Glaser, J. Acfa Chem. Scand. 1979, 33, 7.89.

~~

6648 J. Am. Chem. SOC.,Vol. 103, No. 22, 1981

Glaser and Henriksson

Table V. Chemical Shifts for Various MXm-" Complexes (M = Zn(II), Cd(II), Tl(II1); X = C1, Br, I) in Aqueous Solutions and in Some Solidsh MX(m-I)'

MX,(m-2)+

MX, (m-')+

MX, (m-,

)+

MX5(m-5)+

M x , (m-6).

ref

x = c1 Zn Cd Cd T1

30 89 63 112

295 114 98 115

10 72 62 -5 4 8

511 75 83 -1320

119 292 282 326 - 35g

25 3 495

188=

559

- 64'

162b

22 24 23

-114d

X=Br

Zn Cd Cd

TI

(-230) 365 186 -902

136 379 399 -768

22 24 23 -3357e

X=I Cd Cd T1

47 46

(57) (47)

140 122

71 71 - 3646f

[Co(NH,),l$dC1,. [Co(en),l,[CdC1,1C1,~2H20~Cs,TICI,.H,O. Na3TIC1,~12H,0. e [Co(NH,),]TlBr,. TlC1,.4H,O. Shifts are given in ppm toward higher frequency from M(aq)".

24 23 [NBu,]TII,.

and to the fact that the second coordination sphere is not exactly symmetry is lost when the third cadmium halide complexes are the same. This is in agreement with the fact that tetrahedral VI4formed. The T1Br3(H20)2complex has an essentially trigonalhas been found in the solution on the basis of vibrational spectra290 bipyramidal geometry (both in the solid and in solution), as shown and X-ray diffraction datas3and that slightly distorted tetrahedral by X-ray diffraction r e s u l t ~ and ~ ~ by . ~ similar ~ values of aTIB,,in Tl14- has been found in the solid ~ t a t e . ~ ~ , ~ ~ the solid and in the solution (Table 111). The configuration of The correlation between the chemical shift of the metal nucleus TlC13 in solution is not trigonal bipyramidal but probably apand the geometry of the halide complexes of the dIo ions Zn2+ proximately tetrahedral (vide supra). and Cd2+was discussed by Maciel et al.,22,24who found that the From the data in Table V, some common features can be octahedral complexes are more shielded than the complexes with observed. For complexes with the same geometry, the shielding tetrahedral structure. The shifts for the thallium(II1) chloride of the metal nucleus increases in the order C1-, Br-, I-, Le., with and bromide complexes determined in this work make it possible increasing atomic weight and polarizability and with decreasing to extend this discussion. It would be of great interest to include electronegativity of the halide. The big shifts to lower frequency lg9Hgshifts for the mercury(I1)halide complexes. However, no for TlI4-, T12Br93-,and TlBr63- do probably contain considerable contribution from the "heavy-atom effect" caused by the high such investigation was found in the literature. value of the spin-orbit coupling constant for bromine and iodine.69 The chemical shifts of some halide complexes of the d10ions There are two exceptions to this pattern: ZnX, (X = C1, Br) and Zn2+, Cd2+,and T13+are given in Table V. The structures of the halide complexes in aqueous solutions are probably approxCdBr3 in ref 24. imately the same for Zn(II), Cd(II), and Tl(III), at least as far As pointed out by Maciel et a1.?2 the measurements on the Zn as the octahedral M(H20)6m+,MX(H20)S(m1)+, MXJm6)+, and solutions were performed at [Zn], = 0.5 M, whereas the stability the tetrahedral MX,(m-4)+are c ~ n c e r n e d . ~The ~ ~ MX2~ ~ ~ ~ ~ -constants ~~ used are not necessarily valid for [Zn],, > 0.1 M. (H20)Jm2)+species are probably octahedral for CdS9@and Tl,,l Moreover, the species ZnCl', ZnBr2, and ZnBr3- are never present but ZnX, may be different. Certainly, Raman spectra61 were in the solution to any great extent (at most = 15% of the total interpreted in terms of octahedral ZIIX~(H,O)~,but the unusually zinc concentration). This can lead to some systematic errors in the obtained shifts. On the other hand, if the 67Znshifts for these high values of may indicate that these complexes are not species are of the correct order of magnitude, the difference in octahedral. their changes can be the result of the fact that Zn(I1) is a hard Trigonal-bipyramidal geometry was proposed for ZnC13-, acceptor, whereas Cd(1I) and Tl(II1) are soft.63 ZnBr3- (Raman spectra6'), CdC13-, and CdI< (vibrational spectras9). However, pyramidal CdBr3- is invoked from another The values of aCdBr3-in ref 23 and 24 are quite different. The Raman study.60 In any case, the high entropy values for the different ionic strengths used (4.5 and 3.0 M, respectively) in these investigations cannot be responsible for this discrepancy. As formation of the CdX3- species6, indicate that the octahedral pointed out p r e v i o u ~ l ythe ~ ~values * ~ ~ of the shifts derived for the complexes are only as good as the equilibrium constants used if (53) Glaser, J.; Goggin, P. L.; SandstrBm, M.; Lutsko, V. Acra Chem. .. Scand., in press. the latter are kept constant during the least-squares procedure, (54) Adams, D. M.; Morris, D. M. J . Chem. SOC.A 1968, 695. as is the case for ref 22-24. Thus, the discrepancy is probably ( 5 5 ) Bol. W.: Gerritz. G. J. A,: van Panthaleon van Eck. C. L. J . ADDL .. caused by the use of different sets of equilibrium constants. Crystallogr; 1970, 3, 486. The shifts for the thallium(II1) chloride complexes follow the (56) Caminiti, R.; Johansson, G. Acta Chem. Scand., in press. (57) Maeda, M.; Johansson, G., to be submitted for publication. same pattern as those for CdX,(Z-")+ (X = C1, Br, I). The oc(58) Pocev, S.; Triolo, R.; Johansson, G. Acta Chem. Scand. 1979,33, 179. tahedral complexes are more shielded than the tetrahedral. The (59) Davies, J. E. D.; Long, D. A. J . Chem. Soc. A 1968, 2054. MX3(m3)+complexes have intermediate shielding. Furthermore, (60) (a) Delwaulle, M. L.; Francois, F.; Weimann, J. C.R. Hebd. Seances the difference between the different octahedral complexes is small, Acad. Sci. 1968,206, 186. (b) Delwaulle, M. L. Bull. SOC.Chim. Fr. 1955, 22, 1294. and especially the species with n = 1 and 2 are almost equally (61) Morris, D. F. C.; Short, E. L.; Waters, D. N . J . Inorg. Nucl. Chem. shielded. The ZnX, complexes are, on the other hand, consid1963, 25, 975. erably less shielded than ZnX+ which indicates that they have (62) Christensen, J. J.; Eatough, D. J.; Izatt, R. M. "Handbook of Metal lost the octahedral symmetry. At first sight, the shifts for the Ligand Heats"; Marcel Dekker: New York,1975. (63) (a) Ahrland, S.; Chatt, J.; Davies, N . R. Q.Reu., Chem. Soc. 1958, thallium(II1) bromide complexes seem to behave differently. The 12, 265. (b) Pearson, R. G. J. Am. Chem. Soc. 1963,85, 3533. tetrahedral TlBr4- is, for example, considerably more shielded than (64) Watanabe, T.; Atoji, M. J. Am. Chem. SOC.1950, 72, 3819. the octahedral T1(H20)63+. However, this pattern may be ra(65) Rath, H. J. Thesis. University of Erlangen-Niirnberg, 1972. tionalized as follows. For the four octahedral T1Brn(H20)~$*")+ (66) Glaser, J. Acta Chem. Scand. 1980, 34, 141. (67) Barrowcliffe, T.; Beattie, I. R.; Day, P.; Livingstone, K. J . Chem. Soc. A 1967, 1810. (68) Hoard, J. L.; Goldstein, L. J . Chem. Phys. 1935, 3, 199.

(69) Cheremisin, A. A.; Schastnev, P. V. J . Mugn. Reson. 1980,40,459.

Study of Thallium(III)Halide Complexes complexes (n = 0, 1, 2, and 6), the increase in the shielding when one water molecule is substituted by a bromide is very large and almost constant (-600 ppm). It can now be seen that the tetrahedral TlBri is about 1600 ppm less shielded than a hypothetical octahedral TlBr4(H20)2-complex (which would be -2400 ppm more shielded than Tl(H2O)b3'). If seen this way, the proposed difference in shielding between octahedral and tetrahedral comp l e ~ e iss ~also ~ valid for the Tl(III)-Br- system. A comparison with mercury(I1) halide complexes (which have similar values of the stability constants, the enthalpy and the entropy of formationb2) could reveal whether this is a more general behavior. Stabaty Constants. For the first four bromide complexes there is apparently no significant difference, if the standard deviations are taken into account, between the stability constants obtained for the different total thallium concentrations. The constants are also in good agreement with those given by Ahrland et a1.IoJ1 The existence of a t least one higher complex (n > 4) seems to be significant. For [Tl],, = 0.05 M the 205Tlshift in concentrated HBr solution is 1169 ppm, Le., -150 ppm below the value of 6nBr4-. In 1 M solutions, the corresponding decrease is from 1331 to 1086 ppm, i.e., 225 ppm. It is improbable that 6TlBr4-,which does not change significantly between the different ionic media investigated (cf. Table IV), should change to this great extent and account for the observed shift. However, it was not possible to determine which of the complexes T1Br;- and TlBr2- is formed. In any case, the values of p 6 are very low, in fact much lower than the detection limit given in ref 10. For [Tl],, = 1.0 M and the highest value of R B =~ 11.8, only 10% of the thallium is present as TlBr63-, according to the stability constants in Table 111. The dominating complex is TlBr4- even at very high bromide concentrations. No values for the stability constants for the higher bromide complexes in aqueous solutions were found in the literature. For the first four chloride complexes in the 0.05 M solutions, good agreement is obtained with the data of Ahrland et al. In 1.O M solutions, the TlCl" and TlCl2+complexes are somewhat weaker, p3 is unchanged, and p4 is slightly higher. In 2.6 M solutions, D3 is slightly higher than for the dilute solutions, the other constants being unchanged. However, all the variation of the stability constants for n 5 4 falls within the limit of f 3 standard deviations from the values of ref 10. The existence of higher chloride complexes is evident, but the calculated stability constants should be seen as rough estimates only, as the shifts from solids were used. Thus, in 0.05 M solutions the value of K5 = 0.3 (1) M-l is not significantly higher than the detection limit given by Ahrland et al.lOJ1It is also in agreement with Ks = 0.8 (2) M-' obtained by spectrophotometric meas u r e m e n t ~and ' ~ K5 = 0.3 M-' calculated from extraction data.s In 1 M solutions, both TlC1;- and TlC1;- occur, with formation constants K5 = 0.6 M-I and K4,6= 0.2 W2. Using Raman spectra from approximately 1 M solutions (containing -2 M HC104), Spiro15 claimed that TlC1;- is formed (K4,6= 0.2 M-2) and that for solutions with R I 5, TlC14- and TlC163- accounted for 92-100% of the total thallium concentration. The lower limit of our Ks= 0.2 M-' gives 15% of the total thallium present as TlCl?-. In 2.6 M solutions only TlClb3-occurs, with K 4 6 = 0.8 M-2. It may be noted that the formation constants of TlCl>-and TlCld-

J. Am. Chem. SOC.,Vol. 103, No. 22, 1981 6649 are rather low for all three ionic media. However, a gradual change from only TIC&'- in 0.05 M, through both species in 1 M, to TlC1:- solely in the 2.6 M solutions can be observed. This is understandable if one takes into account that the ratio [H20]/[C1-]fr, varies from 475 to 7 between the 0.05 and the 2.6 M solutions. Even if the chloride is a strong complexing agent for Tl(II1) and strives to substitute the H 2 0molecules in the first coordination sphere of the metal ion, the equilibrium H 2 0 TlClb3- P T1C1s(H20)2- + C1- can be shifted to the right if a sufficient amount of water compared to [Cl-lf,, is present.

+

Concluding Remarks NMR spectroscopy is very suitable for studying the coordination chemistry of Tl(II1). It is possible to obtain information about the stability of the complexes, changes in geometry, and the kinetics of ligand exchange. If the chemical shift is measured at a sufficient number of points, the stability constants can be determined with an accuracy comparable to that obtained from emf studies. As far as formation of weak complexes is concerned, the N M R method is even more sensitive. A comparison of the chemical shifts for the different complexes in solution with the shifts from solid salts with known structure is of great value since it makes it possible to obtain structural information about the complexes in solution. Note Added in Proof. A recent paper by M~Garvey'~ presents some results of direct relevance to our investigation. (1) The chemical shifts for the InC1:- and for the InClb3- species show the same pattern as that observed for and TlClb3-. (2) The chemical shift of the compound I d 3 , which has the iodinebridged structure 121n121n12in nondonor solvents, is close to the shift of the I d 4 - ion. As pointed out, the formal charge apparently has little effect on the resonance frequency, given that both Id4and In& involve In(II1) tetrahedrally bonded to iodine. This fact supports our conclusion about the probable structure of Cs3T12Br6(s). (3) There is no change in the chemical shift or in the line width for the acetonitrile solution of [(C6H5)4P)]InI4,when a threefold excess of [(c&5)$]I is added. This is in agreement with the fact that the maximum coordination number for the I d , , - " complexes is four. These results also show that changes in the second coordination sphere, due to the addition of excess halide, do not influence the 'lsIn chemical shift. Our conclusion on the existence of TlBr:-" species with n > 4 can be seen in this light. Moreover, weak complex formation of InBr?-" (n > 4) is suggested in acetonitrile, whereas both and InClb3- are formed, showing the same trend as the corresponding thallium(II1) species. Acknowledgment. We wish to thank Dr. P. Goggin for a stimulating discussion during the planning of the experiments and Dr. G. Biedermann for help in performing the specific interaction theory calculations. Valuable comments of Professor I. Grenthe are also acknowledged. This work was partially supported by the Swedish Natural Science Research Council (NFR). (70) McGarvey, B. R.; Trudell, C. 0.;Tuck, D. G.; Victoriano, L. Inorg. Chem. 1980, 19, 3432.